Analytical approach:
The authors used a model to synthesise the data from a range of sources to estimate the changes in the percentage time in the therapeutic range (TTR). The time horizon was two years and the authors stated that the perspective was that of the health care payer, which was the Department of Veterans Affairs (VA).
Effectiveness data:
The main sources were identified by a recent systematic review (Singer, et al. 2008, see 'Other Publications of Related Interest' below for bibliographic details) that found 16 randomised controlled trials (RCTs) and two observational studies. The main clinical effectiveness estimates were the Congestive heart failure, Hypertension, Age, Diabetes, and Stroke (CHADS2) score, the percentage TTR, and the risk of adverse events (ischaemic stroke, major haemorrhage, or death) with these two outcomes. The relationship between the TTR and adverse events was from a secondary analysis of a RCT. Poisson meta regression was used to predict the adverse event rates, from the included studies, and the annual event rates for individual patients in the VA population of 67,077 patients. The changes in TTR, with the programme, were assumed and were assumed to be the same for all patients.
Monetary benefit and utility valuations:
The utility values were from a published cost-effectiveness study of patients with atrial fibrillation.
Measure of benefit:
Quality-adjusted life-years (QALYs) were the summary benefit measure.
Cost data:
The cost categories were warfarin therapy (drug and monitoring) and the treatment of adverse events (ischaemic stroke or intracranial haemorrhage). The cost estimates were from published literature and reflected US Medicare costs. All costs were presented in 2008 US dollars ($).
Analysis of uncertainty:
One-way, two-way, and probabilistic sensitivity analyses were performed to assess the impact of parameter uncertainty on the results. The results were presented in cost-effectiveness acceptability curves.